TL;DR
This study used interrupted time series analysis to examine how COVID-19 lockdowns in Victoria, Australia, affected orthopaedic trauma admissions. While the first lockdown saw a significant drop in admissions, subsequent lockdowns had less impact, with a rebound in injuries occurring during periods of eased restrictions, particularly those sustained at home.
What Problem Does This Paper Address?
The COVID-19 pandemic and the associated public health interventions, such as lockdowns, significantly altered daily life globally. While it was anticipated that these measures would reduce trauma incidents, the specific impact on orthopaedic trauma admissions and the characteristics of these injuries across multiple lockdown periods remained unclear. This study aimed to quantify these changes in Victoria, a region that experienced prolonged and multiple lockdowns, to understand the nuanced effects of these interventions on a specific type of injury.
Methods
The researchers employed an interrupted time series analysis (ITSA), a robust statistical method for evaluating the impact of interventions over time. They analyzed weekly data on orthopaedic trauma admissions from the Victorian Orthopaedic Trauma Outcomes Registry (VOTOR) spanning from January 2017 to June 2022. This period included a pre-COVID baseline, multiple lockdown periods, and intermissions between lockdowns. ITSA allowed them to model trends before and after each "interruption" (lockdown or intermission) and assess changes in both the level and slope of admission volumes and injury characteristics. They used Poisson regression for admission volumes and Ordinary Least Squares (OLS) regression for injury event characteristics.
Datasets
The study utilized data from the Victorian Orthopaedic Trauma Outcomes Registry (VOTOR). This registry collects data on adult patients (≥16 years) with emergency orthopaedic admissions of over 24 hours from four hospitals in Victoria, Australia. The dataset included patient demographics, mechanism of injury, management, and in-hospital outcomes. The study period covered from January 2017 to June 2022, encompassing approximately 26,262 orthopaedic trauma admissions.
Key Insights and Ideas
- Lockdown Fatigue is Real: The impact of lockdowns wasn't uniform. The initial lockdown caused a sharp decline, but subsequent ones had less of a dramatic effect, suggesting a potential "pandemic fatigue" or adaptation to restrictions.
- The Rebound Effect: Easing restrictions led to a surge in orthopaedic trauma admissions, exceeding pre-pandemic levels. This highlights the importance of understanding how changes in activity levels post-lockdown influence injury patterns.
- Home is Where the Injuries Are (During Lockdowns): A significant shift occurred towards injuries happening at home during lockdowns, while motor vehicle collisions decreased. This points to changes in daily routines and risk environments.
- Age Matters: Patients admitted during lockdown periods were, on average, older. This could be linked to the decrease in younger-onset injuries like those from motor vehicle accidents.
Key Results
- Admission Volume: The first week of the initial lockdown (L1) saw a 24% decrease in weekly orthopaedic trauma admissions compared to the preceding week. However, the second lockdown (L2) showed admissions approximating pre-COVID levels. Intermission periods and the post-COVID period saw an increase of 33% and 20%, respectively, in weekly admissions compared to the preceding week.
- Injury Location: During lockdowns, there was a higher proportion of injuries occurring at home (38% vs. 28% pre-COVID), and fewer due to motor vehicle collisions (8% vs. 11% pre-COVID).
- Patient Demographics: Patients admitted during lockdown periods were older (mean age 61 vs. 58 years old).
- Injury Characteristics: The pattern of injury by intent, body region injured, injury severity, and in-hospital mortality remained largely unchanged across the periods.
- Activity: A decrease in sports and leisure injuries was observed at the onset of L1, but an increase was noted at the onset of the intermission (I1).
- Length of Stay: Mean length of stay was stable before the pandemic but increased during the intermission and post-COVID periods.
Limitations and Future Work
The authors acknowledge several limitations:
- Patient Allocation: Accurately assigning patients to specific lockdown or intermission periods was complicated by the fact that residential postcodes could belong to multiple local government agencies. This led to a potential for misallocation, with up to 26% of patients in the first lockdown period being incorrectly allocated.
- Aggregation of Periods: For ITS analysis, lockdowns and intermissions had to be aggregated, which might mask finer-grained temporal variations.
- Confounding Factors: The study could not account for broader societal trends like healthcare utilization, socioeconomic status, or patient comorbidities, which could influence injury patterns.
- Time-Varying Confounders: Due to the specific nature of the Victorian COVID-19 situation, a controlled interrupted time series design to address time-varying confounders was not possible.
- Generalizability: As a registry-based study, it may not reflect the experience of all health services in Victoria, and it was not a population-based study.
Future work could explore the impact of other public health measures, investigate confounding factors more deeply, and potentially expand to a national or international comparative analysis. The findings also suggest the need for better planning for staffing and resource allocation during future periods of major disruption.
Summary updated